DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] Enhanced fuel cell performance is obtained through use, in a solid polymer fuel cell electrode, of a carbon-supported catalyst having greater than conventional levels of acidic surface oxide groups thereon. A carbon-supported catalyst pH less than conventional is indicative of these groups. Conventional pH levels appear to be greater than about 5. Herein, pH refers to the value determined by the ASTM standard test method for carbon black, ASTM # D1512-95 Test Method B—sonic slurry test method. This test, which can be applied to carbon-supported catalyst samples, involves measuring the pH of an aqueous slurry containing a carbon black in a specific amount of water. (Note that other methods may be employed to measure the acidity of a carbon-supported catalyst. However, values determined by other methods may not always simply correlate with that of the ASTM method. For instance, a titration method may be employed instead. If the carbon-supported catalyst sample contains residual acid from the preparation process, it may be washed repeatedly in water until no additional acid is removed. The wash water can then be titrated to determine the amount of “free” acid on the carbon-supported catalyst sample. Then, the sample may be subjected to an ion exchange process, in a NaCl solution for example, and the solution may then be titrated to determine the amount of “bound” acid on the carbon-supported catalyst sample. Such a method may distinguish residual acid from the acidic surface oxide groups.)

[0029] A simple way to obtain such a carbon-supported catalyst is to introduce pendant surface oxide groups onto a conventional carbon-supported catalyst by exposing the carbon-supported catalyst to a suitable oxidizing environment. Various oxidizing methods are known, but a chemical treatment using an oxidizing solution is preferred. Various known oxidizing species can be employed in such a solution, including HNO ₃ , H ₃ PO ₄ , KMnO ₄ , KClO ₃ , HF, or (NH ₄ ) ₂ S ₂ O ₈ . However, species which are less hazardous and which are readily available such as HNO ₃ , KMnO ₄ , or (NH ₄ ) ₂ S ₂ O ₈ are preferred. Stronger oxidizing species may be preferable in order to hasten the treatment process. Satisfactory results have been obtained, however, in the Examples described below with both HNO ₃ and the weaker acid H ₃ PO ₄ . (NH ₄ )2S ₂ O ₈ may offer advantages in that it appears to produce stronger acid groups on treated activated carbons than does HNO ₃ , even though HNO ₃ creates larger amounts of oxygen surface complexes. As a further advantage, (NH ₄ ) ₂ S ₂ O ₈ appears not to affect the surface area and pore structure of treated activated carbons as much as HNO ₃ does (see article by C. Moreno-Castilla et al., Carbon, Vol. 35, No. 10-11, p. 1619, 1997).

[0030] The steps involved in chemically treating a conventional carbon-supported catalyst can be relatively simple. For instance, the carbon-supported catalyst may be treated with a strong aqueous acidic solution at an elevated temperature, thereby oxidizing the surface. The preferred concentration of the acid solution, the treatment temperature, and the duration of the treatment will depend on several variables including the acid selected, the nature of the carbon support, and the extent of oxidation required. Some non-inventive empirical trials may be required in order to optimize a treatment for any given carbon-supported catalyst composition for any given application. The parameters that provided satisfactory results in the following Examples also provide guidelines for such optimization. Thus, treatment with strong acidic solutions such as 4 M HNO ₃ at temperatures above about 60° C. for 1 hour or more may be sufficient. Higher concentrations, higher treatment temperatures, and/or longer times may be required when using weaker acids.

[0031] Once the desired oxidation treatment is completed, the carbon-supported catalyst powder is separated from the solution. This can be readily accomplished by washing the powder with water, filtering it, drying the filter cake, and then grinding up the cake to provide a free-flowing powder. HNO ₃ is a preferred choice for the oxidizing species in the treatment solution since it is readily removed afterward.

[0032] In the following, several terms are used to describe certain electrochemical characteristics of carbon-supported catalysts, particularly carbon-supported platinum catalysts. Herein, the electrochemical surface area of the catalyst (ECA) is defined as the catalyst surface area as determined by conventional CO stripping voltammetry in an ex-situ (that is, not in a fuel cell) test configuration. The effective platinum surface area (EPSA) is a dimensionless parameter defined as the catalyst electrochemical surface area divided by the geometric area of the test electrode. EPSA is also determined by CO stripping voltammetry, but is performed in-situ (that is, in a fuel cell). Thus, ECA more closely measures the total catalyst surface area that is accessed by CO while EPSA measures the catalyst surface that is accessed both by CO and a fuel cell electrolyte. The utilization of the catalyst in an electrode is defined as the EPSA divided by (ECA * electrode loading in grams platinum per m ² electrode), which is indicative of the fraction of available catalyst surface that is actually utilized in fuel cell operation.

[0033] For treated carbon-supported platinum catalysts, it has been observed that the EPSA increases as may be expected if the catalyst performance in a fuel cell has been improved. An increase in the measured EPSA is indicative of increased reactant access to the available platinum on the carbon support. However, surprisingly, it was also observed that the ECA may decrease as a result of the treatment. A decrease in ECA is generally indicative of a loss in platinum surface accessible to reactant gases and would generally be assumed to be detrimental and not an advantage. Nonetheless, a significant net improvement associated with treatment has been observed.

[0034] It is expected that the treatment of any carbon-supported catalyst may result in enhanced fuel cell performance, particularly at higher current densities where mass transport issues related to proton conduction are greater. Platinum is a preferred catalyst for use at the cathode of solid polymer fuel cells. The present technique may also be employed to obtain improvements on an anode. However, a common catalyst used in solid polymer fuel cell anodes is Pt—Ru and acid treatments may dissolve the Ru component. Further, present day anode performance is much better than cathode performance and so the greatest fuel cell improvements are expected to come from improvements at the cathode. The carbon support in the carbon-supported catalyst preferably has many sites for the formation of acidic surface oxide groups and is thus preferably a disordered carbon or carbon black, rather than a highly graphitic one. Such carbons are similarly preferred in the making of conventional carbon-supported catalysts by ion exchange methods.

[0035] Instead of forming pendant surface oxide groups on an otherwise conventional carbon-supported catalyst, other alternative methods may also be considered. For instance, conventional carbon-supported catalysts can be prepared by ion exchange methods which involve forming acidic oxide groups on the surface of the carbon support as sites for ion exchange. While these groups are subsequently removed once the introduction of a metal catalyst through ion exchange has been accomplished, an omission or modification of certain process steps may still achieve a desired deposition of an active catalyst without losing the remaining surface groups. For instance, intermediate platinum oxides are generally reduced to platinum metal in commercial carbon-supported catalysts. For use in fuel cell cathodes, which are exposed to an oxidizing environment in the fuel cell, it may be satisfactory to employ the intermediate carbon-supported platinum oxide. If so, the reducing step and loss of oxide groups associated with that step can be avoided. Further, the intermediate carbon-supported platinum oxide may be acceptably reduced in the reducing environment of a fuel cell anode (that is, in situ). If so, the reducing step and loss of oxide groups associated with that step can be avoided can be avoided here too.

[0036] Having obtained a desired carbon-supported catalyst, the fabrication of electrodes and fuel cells using the electrodes can be done in a conventional manner. Typically, fuel cell electrodes are made by applying the carbon-supported catalyst to an electrode substrate (usually a carbon fiber sheet or the like). An ionomer ink is then prepared which comprises an aqueous ionomer solution and the carbon-supported catalyst, and the ionomer ink is applied to a substrate. After drying, portions of the catalyst are desirably in intimate contact with the ion conducting ionomer.

[0037] A fuel cell comprising a treated carbon-supported catalyst is shown in the schematic diagram in FIG. 1 . Fuel cell assembly 10 contains at least one membrane electrode assembly (MEA) comprising a porous cathode 2 and a porous anode 3 that are bonded to a solid polymer membrane electrolyte 4 . Treated carbon-supported catalyst 5 is disposed at the interface between cathode 2 and membrane electrolyte 4 . Ionomer 6 is dispersed over the powder 5 thereby contacting both powder 5 and membrane electrolyte 4 . (The catalyst layer on the anode is not shown.) Oxidant flow field 7 and fuel flow field 8 are pressed against cathode 2 and anode 3 respectively on the faces opposite the membrane electrolyte 4 . Fuel cell assembly 10 has an oxidant inlet 11 , an oxidant outlet 12 , a fuel inlet 13 , and a fuel outlet 14 . Air and hydrogen may, for example, be used as the oxidant and fuel respectively. Electrical power is obtained from the fuel cell at positive and negative terminals 15 and 16 respectively.

[0038] No special requirement is imposed on the materials of construction used in the fuel cell because of the treatment. Unlike certain prior art fuel cells which included “free” sulfuric acid in and around an electrode to enhance proton conduction, the presence of the acidic surface oxide groups on the carbon-supported catalyst does not introduce a corrosion concern with regards to the fuel cell hardware.

[0039] Although not all of the theoretical aspects of the present technique are fully understood, surface treatment of the treated carbon-supported catalysts may lead to improvements for various reasons. Without being bound by theory, one possibility is that the presence of acidic surface oxide groups may greatly increase wetability of the catalyst particles by ionomer ink solution, thereby wetting more catalyst surface and thus improving utilization. Accordingly there may be a net increase observed in the effective catalyst surface area (which is accessible to the reactant, ionically connected to the electrolyte, and electrically connected to the fuel cell current collectors), even if there were a decrease in actual surface area accessible to a reactant. Also, if so, it would presumably be preferable to treat a carbon-supported catalyst before any application to a substrate via an ionomer ink. Another possibility is that even if catalyst coated micro- or meso-pores in the carbon support are not wetted with ionomer, the pores may be wetted with water which becomes proton conducting as a result of the presence of the pendant acidic oxide groups lining the surface of the pores. Again, this may improve catalyst utilization.

[0040] As a further consideration, it may be useful to impregnate the carbon-supported catalyst with acid by vacuum impregnation to more fully access the surfaces. Since it is suspected that the observed performance enhancement may be due, at least in part, to improved proton access to catalyst sites, particularly in pores, further enhancements may also be obtained by improving access of the reactants (for example, oxygen) to those sites. This may be accomplished by impregnating those pores with an impregnate having high oxygen solubility, such as certain fluoro-organic chemicals.

[0041] The following Examples have been included to illustrate different embodiments and aspects of the present technique but these should not be construed as limiting in any way. In the Examples set forth below, the carbon-supported Pt catalysts have been obtained from commercial sources and the preparation method is not known. Regardless of preparation method, the performance improvements noted below clearly cannot be attributed to improved dispersion of deposited Pt by way of the prior art ion exchange mechanism since the Pt catalyst has already been deposited. Another mechanism appears to be involved.

EXAMPLE 1

[0042] Carbon-supported catalyst comprising 10% by weight platinum deposited onto a Shawinigan acetylene black carbon support was obtained from Johnson Matthey. Acidic surface oxide groups were introduced by two different acid treatments. Sample A was treated in 5 M HNO ₃ (100 mL acid solution/g carbon-supported catalyst) at 107° C. for 1 hour. The sample was then rinsed with water, filtered, and dried at 80° C. Sample B was treated in a similar manner except 5 M H ₃ PO ₄ was employed. Sample C was left untreated for comparative purposes.

[0043] Fuel cell testing was then carried out using cathodes prepared using each of the three samples. First, the samples were mixed with Nafion™ ionomer solutions to make ionomer inks. Each ink was then applied to a face of a porous carbon electrode substrate, thereby creating a catalyst layer on the substrate. Anodes for the three fuel cell tests were made in a similar fashion using untreated carbon-supported Pt—Ru catalyst that was applied in a Nafion™ ink to porous carbon anode substrates. Membrane electrode assemblies were then made by bonding test cathodes and anodes onto opposite sides of a solid polymer membrane (catalyst layers facing the membrane).

[0044] Appropriate flow field plates were pressed against the electrodes and other conventional fuel cell hardware was used to complete the fuel cell assemblies.

[0045] The fuel cells were then operated at 80° C. on pure hydrogen and air reactants at the anode and cathode respectively. The cell voltage versus current density characteristics for each fuel cell were determined and the results are shown in FIGS. 2 and 3. The former compares the performance of the HNO ₃ treated sample A against untreated sample C. The latter compares the performance of the H ₃ PO ₄ treated sample B against untreated sample C. Both treated samples show a significant improvement at the higher current densities.

EXAMPLE 2

[0046] The tests carried out in Example 1 were repeated, except in this example the carbon-supported catalyst employed was a commercially available product of Johnson-Matthey comprising 40% by weight platinum deposited on a Vulcan TM furnace black XC72R™ carbon support. Additionally, the samples were treated at room temperature using a vacuum impregnation technique. The samples were pumped down in a vacuum chamber and then a small amount of acid was introduced under vacuum. The chamber was then brought back up to ambient pressure and samples were immediately removed. Finally, samples were rinsed, filtered, and dried at 80° C. as in Example 1. Sample D was treated with 5 M HNO ₃ and sample E was treated with 5 M H ₃ PO ₄ . Sample F was left untreated for comparative purposes.

[0047] In a like manner, cathodes were made using each sample and test fuel cells were made with these cathodes. Again, the fuel cells were operated at 80° C. on pure hydrogen and air and the cell voltage versus current density characteristics for each were determined. FIG. 4 compares the performance of the HNO ₃ treated sample D against untreated sample F. FIG. 5 compares the performance of the H ₃ PO ₄ treated sample E against untreated sample F. In this example, while the HNO ₃ treated sample is significantly better than the untreated sample, the H ₃ PO ₄ treated sample does not show much improvement. At the highest current density tested, the H ₃ PO ₄ treated sample E appears slightly worse than the untreated sample, perhaps indicative of a flooding problem. Thus, it appears that this room temperature impregnation treatment was insufficient to successfully treat this specific carbon-supported catalyst with H ₃ PO ₄ .

EXAMPLE 3

[0048] Testing similar to Example 2 was carried out, except that the carbon-supported catalyst employed was a commercially available product of Johnson-Matthey comprising 20% by weight platinum on Vulcan furnace black XC72R™ carbon support. Sample G was treated this time with 4 M HNO ₃ (20 mL acid solution/g carbon-supported catalyst) at 60° C. for 1 hour. Sample H was left untreated for comparative purposes.

[0049] Again, cathodes were made using each sample and test fuel cells were made and operated as in the preceding Examples. FIG. 6 compares the performance of the treated sample G against untreated sample H. The treated sample shows a substantial improvement over that of the untreated sample.

[0050] Some additional characteristics of the treated and untreated samples were determined in this Example. The pore size distributions of each sample were determined by mercury intrusion. FIG. 7 shows the pore size distributions expressed as rate of change of pore volume with log pore diameter versus pore diameter.

[0051] The pH, electrochemical surface area of the Pt (ECA), effective platinum surface area (EPSA) and utilization characteristics were also determined for the treated and untreated samples and the results are shown in the Table below. The pH was determined by the aforementioned ASTM test. ECA was determined by electrochemically stripping adsorbed CO from the Pt surface from a sample mounted on gold gauze in 0.5 M sulfuric acid electrolyte. EPSA was determined by CO stripping voltammetry on a fuel cell electrode sample using 1% CO in nitrogen as the cathode gas. The utilization was then calculated by dividing the EPSA by (ECA * electrode loading in grams Pt per m ² electrode) as mentioned above. 1

[0052] As expected, the pH of the carbon-supported catalyst was reduced substantially by the acid treatment. Also, as expected, the EPSA was substantially increased by the acid treatment. Unexpectedly, however, it appears that the ECA was reduced by the treatment. The increase in utilization is actually much greater than expected since the ECA has decreased significantly, suggesting some loss of Pt surface. Nonetheless, a net benefit was observed in a fuel cell application.

[0053] While particular elements, embodiments and applications of the present invention have been shown and described, it will be understood, of course, that the invention is not limited thereto since modifications may be made by those skilled in the art without departing from the spirit and scope of the present disclosure, particularly in light of the foregoing teachings.